Transcript ENSEMBLES the climate observation dataset E-OBS Albert Klein Tank, KNMI ENSEMBLES together with UEA, Univ.Oxford, MeteoSwiss.

ENSEMBLES
the climate observation dataset
E-OBS
Albert Klein Tank, KNMI
ENSEMBLES
together with UEA, Univ.Oxford, MeteoSwiss
outline
 motivation
 scope
 method
 results
 applications
 outlook
motivation I
ensemble median bias in Q95 of daily Tmax
 RCM evaluation
Summer
at the time of
PRUDENCE
evaluation of
control climate
mainly against
CRU-monthly
grids or station
point data
Kjellstrom et al., Climatic Change, 2007
motivation II
station trends in winter precipitation
 climate change
assessments EEA
scope
 E-OBS
variables: Temp (min, max, mean) and Precip
area: Europe (including the Mediterranean)
resolution: down to 25km (matching RCM grids)
time step: daily (particular focus on extremes)
period: 1950-now
method I
data availability for E-OBS (version 2.0)
 use is made of long
daily station series
from ECA&D and
other EU-projects
Klok and Klein Tank, Int. J. Climatol., 2008
method II
 statistical tests for
number of breaks
in each series
Begert et al.,
Meteor. Z., 2008
method III
 two-step interpolation:
1. monthly means/totals using thin-plate splines
to define the underlying spatial ‘‘trend’’,
taking into account the station elevation
2. daily anomalies using global kriging
with a single variogram for all days, a search
radius of ~500 km and incorporating elevation
dependencies of temperature
Haylock et al., J. Geophys. Res., 2008
method IV
cross validation exercise , e.g. daily Tmean anom.
 global kriging:
most accurate for
daily anomalies
Hofstra et al., J. Geophys. Res., 2008
results I
 output
consists of
daily fields
for each
variable
+ standard
error
(= interpolation
error only)
results II
…but interpolation leads to reduction of extremes
 E-OBS provides
grid box averages
 interpolation is
done to a higher
resolution master
grid, from which
lower resolution
products are
derived
0.1° master
grid
temperature
precipitation
0.25°, 0.50° and
25km, 50km products
results III
 E-OBS is available from: eca.knmi.nl/ensembles
applications I
 evaluation of ERA40-driven RCMs over the Alps:
bias in Q90 of daily precipitation, SON, 1961-90
Pall et al., in prep.
applications II
 evaluation of trends (1961-2000)
in Q05 of Tmin
˚C/decade
CHMI-Aladin RCM
E-OBS
Lister and Jones, in prep.
applications III
 projected
evaluationchange
of precipitation
in precipitation
GEVs in
extremes
Rhine basin
Winter 5-day extremes
Summer 1-day extremes
Hanel & Buishand, submitted to Clim. Dyn., 2009
applications IV
RCMs
mm
extreme
1-day
precipitation
which occurs
once in
5-10 yr
(based on
1950-2008)
Winter
Spring
Summer
Autumn
E-OBS
Lenderink, submitted to Clim. Res., 2009
outlook
 E-OBS was essentially constructed for RCM
evaluation and climate change assessment, but
is one of the lasting outcomes of EMSEMBLES
 E-OBS has spawned similar dataset
developments in Mexico, South America, etc.
 Complementary apects of E-OBS and regional
reanalysis will be used for near-real time
monitoring of daily extremes across Europe in
EURO4M (new EU-FP7 project, 2010-2013)
thank you!
WP5.1 team:
ENSEMBLES
UEA, Univ.Oxford, MeteoSwiss, KNMI